Sunday, January 14th 2018
Intel Could Ditch AMD dGPU Die on Future Core G-series MCMs with "Arctic Sound"
Intel did the impossible in 2017, by collaborating with rival AMD after decades, on a product. The new Core i7-8000G series processors are multi-chip modules that combine quad-core "Kaby Lake" CPU dies with discrete AMD Radeon Vega GPU dies that have their own dedicated HBM2 stacks. With performance-segment notebooks and sleek AIO desktops building momentum for such products, Intel sees a future in building its own discrete GPUs, at least dies that can replace the AMD Radeon IP from its Core G-series processors.
With former AMD Graphics head Raja Koduri switching to Intel amidst rumors of the company investing in discrete GPUs of its own, details emerge of the company's future "Arctic Sound" and "Jupiter Sound" graphics IP, which point to the possibility of them being discrete GPU dies based on the Gen 12 and Gen 13 graphics architectures, respectively. According to Ashraf Eassa, a technology stock commentator with "The Motley Fool," both "Arctic Sound" and "Jupiter Sound" are discrete GPU dies that connect with Intel processor dies over EMIB, the company's proprietary high-density interconnect for multi-chip modules. It could be a long wait leading up to the two, since the company is still monetizing its Gen 9.5 architecture on 8th generation Core processors.
With former AMD Graphics head Raja Koduri switching to Intel amidst rumors of the company investing in discrete GPUs of its own, details emerge of the company's future "Arctic Sound" and "Jupiter Sound" graphics IP, which point to the possibility of them being discrete GPU dies based on the Gen 12 and Gen 13 graphics architectures, respectively. According to Ashraf Eassa, a technology stock commentator with "The Motley Fool," both "Arctic Sound" and "Jupiter Sound" are discrete GPU dies that connect with Intel processor dies over EMIB, the company's proprietary high-density interconnect for multi-chip modules. It could be a long wait leading up to the two, since the company is still monetizing its Gen 9.5 architecture on 8th generation Core processors.
45 Comments on Intel Could Ditch AMD dGPU Die on Future Core G-series MCMs with "Arctic Sound"
Samsung is not a single company. It's a group. It does many things, but withing different companies. When Samsung Electronics needs a metal shell for their phone, it can choose a Samsung Group subsidiary as a supplier, but it has to pay just like any other client.
@notb i never took it personally why would I , im not the one with shares in a company continuously bad mouthing it's main competitor at every opportunity like it's a job, I made a missive comment based largely on delivery times of intel dGpu.
And noted your condescension of Amd , stop being an arse and ill stop pointing out to the people your trying to inform ,just how unvalid and biased your Bs is.
1) No, Samsung is not self sustainable. They would have to join all subsidiaries, which would result in a mess (I don't even think it's possible).
2) No, margins aren't larger. They sell for the same price.
3) And, most importantly, NO, there is no explicit "parent company". There's a net of dependencies, which usually results in Lee family controlling each company. But they can't decide on their own.
For example: Lee family members and other Samsung companies have 25% of Samsung Electronics, so they're in full control in stable times, but if they were going too far, the rest of investors could easily oppose them. And of course Lee's don't control 100% these other S. companies, but they have enough votes to be able to choose the representative for Samsung Electronics general meetings. That's how a conglomerate works.
I've found a graph from 2014 (a simplified version: only the major companies and share percentages are displayed).
Realistically , they are probably one of the least capable of taking over the GPU market within any reasonable time frame. They are just way behind.
I find it hilarious how everyone thought that Raja single handedly ruined RTG but now he will supposedly make miracles at Intel. And both of those assumption are wrong I would add.
Performance-oriented Iris Pro chips have surpassed pre-Vega APUs quite a while ago.
The cheap HD IGPs aren't far behind. HD630 offers 50-70% of A12-9800 performance in games (e.g.
And remember that HD chips are not optimized for performance. It's a cheap IGP with surprisingly decent features that's meant to be put in every consumer CPU.
As for dGPUs or something like Vega IGP - no, Intel doesn't know how to make them as well as NVIDIA and AMD - they've never really bothered about this niche. Gamers will buy dGPUs anyway.
They got interested lately - mostly because of how much software uses GPU acceleration, so Intel CPUs started to lose ground in so called "productivity".
Another reason is how small NVIDIA mobile GPUs are becoming. Intel had a monopoly for graphics in thin notebooks. Now you can buy an ultrabook with MX150. They needed an answer. For the time being it'll be Vega. We'll see what comes next.
And I'm still a believer in Radeon takeover. :)
You should not evaluate Intel's potential by the fact that they don't make discrete gaming GPUs. And the IGPs (both HD and Iris) are top notch.
Also, I don't know what you have against Xeon Phi. It's a very powerful chip. It's just slightly late to the game.
Xeon Phis are doing very well in HPC market.
It could find its way into consumer PCs as well.
If Intel came up with this idea 5 years earlier, I doubt science and enginnering would be so dominated by CUDA today. Everyone would use x86 coprocessors - they're just way easier to use.
Knights ferry started as a gpu , did you not know that , it's hpc use is both niche and circumstantial ie they researched it (for gfx ,made it and it was still shit ,after said Research) so yeah they're going to try and sell it , obviously.
And feck the ifs , arm would've not existed if intel had the foresight you wish they had.........
The idea of many-core processors is getting a lot of traction in the top500 list.
29 computers use Xeon Phi (20 in top100)
by comparison:
87 computers use NVIDIA Tesla (19 in top100)
Not bad, right?
And there are a few other many-core contenders (Sunway, PEZY etc).
It's a very good idea in general. It's just that host CPU-only solutions are easier to code for and CUDA got very, very popular. But many-cores will catch up.
And you wouldn't call super computer's niche.
And the science , research and database crunching they do is typical use case.
I don't search Google before every ,or any reply ,i just remember what I've seen and done and some workloads won't be accelerated by more cpu cores ,shit you and a million others are all about ipc and more cores being a waste in any other thread but here more is better, no the right core for the job ,ie graphics will always be better then a great many x86 cores on a chip ,intel already provided the proof about three times that i actually lived through ,i remember intel poo pooing add in card graphics in the pentium era and a few times since ,only once with an actual gpu though ,their Hd era igp , so in short bring it intel.
Don't talk about it 3-5 years early , just effin do it, i would love three discreet gpu makers ,please do bring it , but till then stfu intel ,and pay your damn bills on time to Amd not like your throwback compo payments, i want them healthy too , all the big pc three healthy and wealthy and INNOVATIVE.
Quit the pr job for intel dude your crap at it.
So this is weird for outsiders, but not unheard of in Korea. They have more examples of chaebols: Hyundai, SK Group, LG, Lotte and so on. The economy is run by ~10 or so large groups.
Not so long ago Daewoo was the second largest group in Korea, but they've went bankrupt. In fact many chaebols went down in 90s. It's not the best business concept.
For some reason Korean groups like to invest in Poland, so I've been interested in the problem for some time.
It started with the now-defunct Daewoo...